For all of our gaming tests we will be testing at both 1920×1080 as well as 1280×1024 resolutions. As of August, 2011, 1920×1080 is the most popular resolution used by gamers according to the Steam Hardware Survey, with nearly 23% of PC gamers using this resolution. http://store.steampowered.com/hwsurvey. However, according to the survey, 24% of gamers still play at 1280×1024 or below, so we will be testing 1280×1024 as well to demonstrate a more CPU-bound scenario.
In general, we test with all detail options at their highest in-game settings in order to stress our hardware to the fullest extent possible. Ideally, our two SLI GTX470 video cards with their 1280MB of DDR5 will be able to handle the pixel pushing duties, isolating our CPU and memory subsystems in our benchmarks. Our i7 2600K CPU is overclocked to 4.4GHz, which should supply sufficient processing power for most of today’s gaming scenarios.
Our first game test involves Crysis 2. Despite the fact that many PC gamers wrote off and ignored the game shortly after release for a variety of reasons, Crysis 2 has been patched and updated several times, making it one of the best looking games on the PC. For our tests we are using the latest version of the game, which is 1.9 as of the time of this writing. We also patched Crysis 2 using the High Resolution Texture Pack as well as the DirectX 11 update. We are using the highest available in-game video quality settings and the DirectX 11 rendering path at two different resolutions in order to place as much strain as possible on the hardware.
Our benchmark takes place during the “Second Chance” sequence of the game. During the benchmark run, the player uses the binoculars to scout the combat area before he engages in several firefights with enemy soldiers in an open outdoor environment. The sequence involves several explosions and both gunfire and melee action. We will be measuring our results using Fraps and will be reporting both minimum and average frames per second (FPS).
The results show minor improvements in average FPS at both resolutions. The minimum FPS at the lower resolution is unchanged while a 1 FPS increase in minimum FPS at 1920×1080 results in a numerically significant 4.6% increase in reported speed, but it is unlikely the user would notice a 1 FPS improvement in actual use.